So, how to do this? One can start by making use of the "sense checker." If your genealogy program has one. It will notify you of births before or after a certain age, marriage at a certain age, numbers of children or spouses etc. Just being aware of such discrepencies might set you straight. One should always endeavor to make sure that things are written or copied in an orderly and sensible manner, even if the source itself in not so arranged.
As an example I was researching some old German church records in Thuringia from the late 1500s and early 1600s. Beside the hurdles of the terrible handwriting and excessive use of abbreviations whose true meaning was known only to that pastor, the microfilmed copies were filmed in a hodge podge manner. That is, they were all the left pages and then all the right pages. So far so good, but unfortunately the pages mostly did not have page numbers, and the guide items like "here starts the births for years" were lacking. In other words, the source itself had been put together in a strange and time consuming way. That series of church records took an enormous time to get through and sort out.
So, someone who might have gone through and not been as careful as I tried to be might have gotten births, marriages and deaths all mixed up because of the commonality of names in a small town, and the mixture of vital records requiring close scrutiny. If I had been a "rush-to-publish Ralph," further users might well have seen the kind of child births or aged births referenced above. After you read and extract a record, think about it! Just because it's written doesn't make it so.
We all know what happens when an obit is put in the paper by someone not terribly concerned with accuracy. Just think of the stories in your own families that you know the truth about but had to be shortened to fit into a death notice. Most people do not have fully written obituaries. They have brief death notices. "Everyone knows" the truth about such and such a spouse or child – but maybe not. Were they stepchildren? Was it the first marriage?
Another true story. About 30 years ago three nice women of a certain age came into my library looking for family information. As it turns out, they were part of what I like to call "greater family" or "suburban family" meaning that those terms are more politely correct than "shirt tail relation." (That latter term has sometimes been defined as those who hang on to somebody's shirt tail, or are dependent on somebody to hitch a ride to fame, fortune, etc. by coasting along in somebody else's wake, or even worse, a woodpile relative - meaning you came from an assignation outside marriage - behind the woodpile.)
To get back on point, I told them that I had information on their family and that they were welcome to read it. They did, and said "Oh, no, A was our mother, not B. But father caught A cheating and banished her away from the city, and married B. We have always thought of B as our natural mother, and not the woman who actually gave birth to us." These women were of such an age that they had all been born between censuses, and so when Mr. So and So was single he was in the first census; then got married and had these three daughters with A; then divorced A and married B; and by the time of the next census he's 1) married, 2) has three children, and 3) the relation of those children is given to the head of household – the father, and not their step-mother. The stepmother appears as wife to the man, which is normal – she is. This was years before censuses became indexed, or digitally available online an so forth. It was a natural mistake on my part to see B's obit which said "wife of the deceased father, and mother of C, D, and E," the daughters, who were really stepdaughters.
See why you have to use sense checkers and common sense as well?
Also, be sure to get the right name and location consistently. There's nothing wrong with places being renamed, or people being known by different names. It happens all the time. But do be careful that you proofread your materials. For example, if you are researching a name, especially a common one, make sure that you are open to variant spellings. Just today I had a patron who was working on the name Clark. He was absolutely sure that they had never used Clarke with an e. (Funny thing – the local volunteer who was helping him was a Clarke with an e). But various sources in the collection used both spellings for the correct family.
There was a case recently on a message board where the immigrant to the U.S. used several names. If you stick to just one version you may well miss useful information under another variant. That may seem pretty obvious, but sometimes it is not, especially to those who have not been searching family for very long. This particular name ended in -O, but also appeared as – IO. And the first name appeared and Daniel, Dana, Danna, Dannell, Demetrio (!), and Demetro. His wife's name was Lena. Or Lina. Or Magdalena. Or Magda. Or Magdelena. And a relative was named Mabe. Or Mabee. Or Maby. Or Mabie.
Or maybe I am making too much of this, but perhaps not (pun intended).
You not only have to scan a wide area geographically, but be open to a lot of different spellings.
Be advised that some things may be true. I just today read and article about a man who turned 113. That would make him born in 1896. Lo and behold, there he is as a child in the Minnesota census for 1900. And in 2009 CNN interviewed him and he gave data that talked about his whole life. (I personally found interesting that he did not retire until 99, but that's my own prejudice).
And pretty please, cite your sources. There are many ways to learn how to do that. Simply claiming "I said so" is not enough, but many people do research that way.
Elizabeth Shown Mills is a recognized genealogical expert and I recently found a quote in a review of her book, Evidence!, that reads:
"Successful research--research that yields correct information with a minimum of wasted time and funds--depends upon a sound analysis of evidence. Source citation is fundamental, but it is not enough. The validity of any piece of evidence cannot be analyzed if its source is unknown. Citing a worthless source is an effort that produces worthless results. Research, evidence, citation, and analysis are inseparable. Evidence is the vehicle that moves our research from curiosity to reality. Citation and analysis are the twin highways that get us there, smoothly and safely."
This is another important bit of advice that is often overlooked. Being a genealogical librarian, I can see what people do when they perform research. If we are lucky they make a copy and take it with them. Often they just write something down on a piece of paper and come back weeks later and cannot remember where they found that information. Well, we are good, but we are not mind readers. Guessing where in 40,000 books they found something is not a possibility that happens often.
Cite your sources! Think things through! Use common sense! Sound practices are easily overlooked (guilty here as well!) but they are very important.